Towards an efficient and interpretable Machine Learning approach for Energy Prediction in Industrial Buildings: A case study in the Steel Industry

Energy consumption worldwide has increased significantly over the past few decades due to increasing population and economic growth. Efficient building energy consumption prediction plays an important role in energy planning, management, and conservation in smart buildings. This paper presents a new machine learning (ML) approach for the prediction of energy consumption in industrial buildings. It presents a trade-off between the performance and the interpretability of ML models which are major issues for building energy prediction using ML. The applicability of the proposed approach is demonstrated by a real case study to predict energy consumption in the steel industry. It shows that the Random Forest (RF) model provides the most effective prediction results, and the permutation feature importance helps the steel industry experts to better understand the judgments of the model. Hence, it will assist them to optimize energy consumption and make the most appropriate decision in industrial buildings.